The Developmental Arc of Face Recognition
Face recognition ability develops through childhood and peaks in young adulthood, typically in the late teens to mid-20s. After peaking, performance on face recognition tests (especially for unfamiliar faces and in challenging conditions) gradually declines through the 40s and more rapidly from the 60s onward.
The decline is selective: recognition of very familiar faces (close family members, long-term friends) remains well-preserved into advanced age. The deficit affects primarily unfamiliar face matching and learning of new faces, exactly the tasks where aging disrupts the encoding efficiency of the fusiform face area and the working memory resources needed for effortful comparison.
AI Performance Is Age-Independent
Unlike human face recognition, AI face recognition does not degrade with the age of the operator, the system is equally accurate at any time, without fatigue or attentional variation across a session. This is one of the structural advantages of AI for tasks that require consistent face comparison performance over extended periods.
The interaction between human and AI aging is interesting in the context of celebrity databases. Older users matching against older celebrities share an age-related appearance range that is well-represented in databases spanning many decades. The AI handles cross-age matching better than human observers in general.